Interviewing people who are being audited

Anxiety produces the answer they think you want.

People being audited want to give the right answer and will construct one. This is not evasion; it is ordinary social behaviour under evaluation, and it systematically corrupts audit evidence unless the interviewer works against it.

Open questions about actual instances work better than questions about process. Asking somebody to describe the last time they did this produces a sequence with the exceptions in it; asking how the process works produces the procedure recited.

Following the artefact is the most reliable technique. Pick a real job, order or record and trace it through every hand it passed, asking each person what they did with it. Discrepancies appear without anybody being asked to admit anything.

Silence is a tool. People fill it, and what they add after a pause is frequently the qualification that matters, which they would not have volunteered in answer to a direct question.

The interviewee should never be surprised at the closing meeting by something they said. Confirming an understanding at the time, in their words, is both fair and produces better evidence.

It is worth adding that interviewing the newest person in a team is disproportionately informative. They know what they were actually taught rather than what the procedure says, and they have not yet learned which questions to answer carefully.